Features

• IELTS Writing (Task 1 / Task 2): rubric scoring, error analysis, improvement suggestions, revised samples.
• RAG-based few-shot retrieval to provide grounded feedback.
• Optional Obsidian sync for saving scoring records and metadata.
• Image word cards: point/box interactive segmentation, auto English naming, Chinese translation.
• Dictionary IPA/pronunciation/examples (fallback to Qwen examples when missing).
• Word-card CRUD (view / edit / delete) and dark/light theme.

Local Run (Quick Start)

Install dependencies:

pip install -r requirements.txt

Copy .env.template to .env and fill your keys (e.g. Qwen API Key, HF Token). Then initialize data/indexes:

python scripts/setup_data_and_indexes.py

Start the server:

python app.py

Open: http://127.0.0.1:5000/ (home) and http://127.0.0.1:5000/word-learning (word learning).